1 /* vi: set sw=4 ts=4: */
3 * wget - retrieve a file using HTTP or FTP
5 * Chip Rosenthal Covad Communications <chip@laserlink.net>
6 * Licensed under GPLv2, see file LICENSE in this source tree.
8 * Copyright (C) 2010 Bradley M. Kuhn <bkuhn@ebb.org>
9 * Kuhn's copyrights are licensed GPLv2-or-later. File as a whole remains GPLv2.
12 //usage:#define wget_trivial_usage
13 //usage: IF_FEATURE_WGET_LONG_OPTIONS(
14 //usage: "[-c|--continue] [-s|--spider] [-q|--quiet] [-O|--output-document FILE]\n"
15 //usage: " [--header 'header: value'] [-Y|--proxy on/off] [-P DIR]\n"
16 //usage: " [--no-check-certificate] [-U|--user-agent AGENT]"
17 //usage: IF_FEATURE_WGET_TIMEOUT(" [-T SEC]") " URL..."
19 //usage: IF_NOT_FEATURE_WGET_LONG_OPTIONS(
20 //usage: "[-csq] [-O FILE] [-Y on/off] [-P DIR] [-U AGENT]"
21 //usage: IF_FEATURE_WGET_TIMEOUT(" [-T SEC]") " URL..."
23 //usage:#define wget_full_usage "\n\n"
24 //usage: "Retrieve files via HTTP or FTP\n"
25 //usage: "\n -s Spider mode - only check file existence"
26 //usage: "\n -c Continue retrieval of aborted transfer"
27 //usage: "\n -q Quiet"
28 //usage: "\n -P DIR Save to DIR (default .)"
29 //usage: IF_FEATURE_WGET_TIMEOUT(
30 //usage: "\n -T SEC Network read timeout is SEC seconds"
32 //usage: "\n -O FILE Save to FILE ('-' for stdout)"
33 //usage: "\n -U STR Use STR for User-Agent header"
34 //usage: "\n -Y Use proxy ('on' or 'off')"
38 //#define log_io(...) bb_error_msg(__VA_ARGS__)
39 #define log_io(...) ((void)0)
54 off_t content_len
; /* Content-length of the file */
55 off_t beg_range
; /* Range at which continue begins */
56 #if ENABLE_FEATURE_WGET_STATUSBAR
57 off_t transferred
; /* Number of bytes transferred so far */
58 const char *curfile
; /* Name of current file being transferred */
62 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
66 char *fname_out
; /* where to direct output (-O) */
67 const char *proxy_flag
; /* Use proxies if env vars are set */
68 const char *user_agent
; /* "User-Agent" header field */
69 #if ENABLE_FEATURE_WGET_TIMEOUT
70 unsigned timeout_seconds
;
74 smallint chunked
; /* chunked transfer encoding */
75 smallint got_clen
; /* got content-length: from server */
76 /* Local downloads do benefit from big buffer.
77 * With 512 byte buffer, it was measured to be
78 * an order of magnitude slower than with big one.
80 uint64_t just_to_align_next_member
;
81 char wget_buf
[CONFIG_FEATURE_COPYBUF_KB
*1024];
83 #define G (*ptr_to_globals)
84 #define INIT_G() do { \
85 SET_PTR_TO_GLOBALS(xzalloc(sizeof(G))); \
86 IF_FEATURE_WGET_TIMEOUT(G.timeout_seconds = 900;) \
90 /* Must match option string! */
92 WGET_OPT_CONTINUE
= (1 << 0),
93 WGET_OPT_SPIDER
= (1 << 1),
94 WGET_OPT_QUIET
= (1 << 2),
95 WGET_OPT_OUTNAME
= (1 << 3),
96 WGET_OPT_PREFIX
= (1 << 4),
97 WGET_OPT_PROXY
= (1 << 5),
98 WGET_OPT_USER_AGENT
= (1 << 6),
99 WGET_OPT_NETWORK_READ_TIMEOUT
= (1 << 7),
100 WGET_OPT_RETRIES
= (1 << 8),
101 WGET_OPT_PASSIVE
= (1 << 9),
102 WGET_OPT_HEADER
= (1 << 10) * ENABLE_FEATURE_WGET_LONG_OPTIONS
,
103 WGET_OPT_POST_DATA
= (1 << 11) * ENABLE_FEATURE_WGET_LONG_OPTIONS
,
111 #if ENABLE_FEATURE_WGET_STATUSBAR
112 static void progress_meter(int flag
)
114 if (option_mask32
& WGET_OPT_QUIET
)
117 if (flag
== PROGRESS_START
)
118 bb_progress_init(&G
.pmt
, G
.curfile
);
120 bb_progress_update(&G
.pmt
,
123 (G
.chunked
|| !G
.got_clen
) ? 0 : G
.beg_range
+ G
.transferred
+ G
.content_len
126 if (flag
== PROGRESS_END
) {
127 bb_progress_free(&G
.pmt
);
128 bb_putchar_stderr('\n');
133 static ALWAYS_INLINE
void progress_meter(int flag UNUSED_PARAM
) { }
137 /* IPv6 knows scoped address types i.e. link and site local addresses. Link
138 * local addresses can have a scope identifier to specify the
139 * interface/link an address is valid on (e.g. fe80::1%eth0). This scope
140 * identifier is only valid on a single node.
142 * RFC 4007 says that the scope identifier MUST NOT be sent across the wire,
143 * unless all nodes agree on the semantic. Apache e.g. regards zone identifiers
144 * in the Host header as invalid requests, see
145 * https://issues.apache.org/bugzilla/show_bug.cgi?id=35122
147 static void strip_ipv6_scope_id(char *host
)
151 /* bbox wget actually handles IPv6 addresses without [], like
152 * wget "http://::1/xxx", but this is not standard.
153 * To save code, _here_ we do not support it. */
156 return; /* not IPv6 */
158 scope
= strchr(host
, '%');
162 /* Remove the IPv6 zone identifier from the host address */
163 cp
= strchr(host
, ']');
164 if (!cp
|| (cp
[1] != ':' && cp
[1] != '\0')) {
165 /* malformed address (not "[xx]:nn" or "[xx]") */
169 /* cp points to "]...", scope points to "%eth0]..." */
170 overlapping_strcpy(scope
, cp
);
173 #if ENABLE_FEATURE_WGET_AUTHENTICATION
174 /* Base64-encode character string. */
175 static char *base64enc(const char *str
)
177 unsigned len
= strlen(str
);
178 if (len
> sizeof(G
.wget_buf
)/4*3 - 10) /* paranoia */
179 len
= sizeof(G
.wget_buf
)/4*3 - 10;
180 bb_uuencode(G
.wget_buf
, str
, len
, bb_uuenc_tbl_base64
);
185 static char* sanitize_string(char *s
)
187 unsigned char *p
= (void *) s
;
194 static FILE *open_socket(len_and_sockaddr
*lsa
)
198 /* glibc 2.4 seems to try seeking on it - ??! */
199 /* hopefully it understands what ESPIPE means... */
200 fp
= fdopen(xconnect_stream(lsa
), "r+");
202 bb_perror_msg_and_die(bb_msg_memory_exhausted
);
207 /* Returns '\n' if it was seen, else '\0'. Trims at first '\r' or '\n' */
208 static char fgets_and_trim(FILE *fp
)
213 if (fgets(G
.wget_buf
, sizeof(G
.wget_buf
) - 1, fp
) == NULL
)
214 bb_perror_msg_and_die("error getting response");
216 buf_ptr
= strchrnul(G
.wget_buf
, '\n');
219 buf_ptr
= strchrnul(G
.wget_buf
, '\r');
222 log_io("< %s", G
.wget_buf
);
227 static int ftpcmd(const char *s1
, const char *s2
, FILE *fp
)
233 fprintf(fp
, "%s%s\r\n", s1
, s2
);
235 log_io("> %s%s", s1
, s2
);
240 } while (!isdigit(G
.wget_buf
[0]) || G
.wget_buf
[3] != ' ');
242 G
.wget_buf
[3] = '\0';
243 result
= xatoi_positive(G
.wget_buf
);
248 static void parse_url(const char *src_url
, struct host_info
*h
)
253 h
->allocated
= url
= xstrdup(src_url
);
255 if (strncmp(url
, "http://", 7) == 0) {
256 h
->port
= bb_lookup_port("http", "tcp", 80);
259 } else if (strncmp(url
, "ftp://", 6) == 0) {
260 h
->port
= bb_lookup_port("ftp", "tcp", 21);
264 bb_error_msg_and_die("not an http or ftp url: %s", sanitize_string(url
));
267 // "Real" wget 'http://busybox.net?var=a/b' sends this request:
268 // 'GET /?var=a/b HTTP 1.0'
269 // and saves 'index.html?var=a%2Fb' (we save 'b')
270 // wget 'http://busybox.net?login=john@doe':
271 // request: 'GET /?login=john@doe HTTP/1.0'
272 // saves: 'index.html?login=john@doe' (we save '?login=john@doe')
273 // wget 'http://busybox.net#test/test':
274 // request: 'GET / HTTP/1.0'
275 // saves: 'index.html' (we save 'test')
277 // We also don't add unique .N suffix if file exists...
278 sp
= strchr(h
->host
, '/');
279 p
= strchr(h
->host
, '?'); if (!sp
|| (p
&& sp
> p
)) sp
= p
;
280 p
= strchr(h
->host
, '#'); if (!sp
|| (p
&& sp
> p
)) sp
= p
;
283 } else if (*sp
== '/') {
286 } else { // '#' or '?'
287 // http://busybox.net?login=john@doe is a valid URL
288 // memmove converts to:
289 // http:/busybox.nett?login=john@doe...
290 memmove(h
->host
- 1, h
->host
, sp
- h
->host
);
296 // We used to set h->user to NULL here, but this interferes
297 // with handling of code 302 ("object was moved")
299 sp
= strrchr(h
->host
, '@');
301 // URL-decode "user:password" string before base64-encoding:
302 // wget http://test:my%20pass@example.com should send
303 // Authorization: Basic dGVzdDpteSBwYXNz
304 // which decodes to "test:my pass".
305 // Standard wget and curl do this too.
307 h
->user
= percent_decode_in_place(h
->host
, /*strict:*/ 0);
314 static char *gethdr(FILE *fp
)
321 /* retrieve header line */
322 c
= fgets_and_trim(fp
);
324 /* end of the headers? */
325 if (G
.wget_buf
[0] == '\0')
328 /* convert the header name to lower case */
329 for (s
= G
.wget_buf
; isalnum(*s
) || *s
== '-' || *s
== '.'; ++s
) {
330 /* tolower for "A-Z", no-op for "0-9a-z-." */
334 /* verify we are at the end of the header name */
336 bb_error_msg_and_die("bad header line: %s", sanitize_string(G
.wget_buf
));
338 /* locate the start of the header value */
340 hdrval
= skip_whitespace(s
);
343 /* Rats! The buffer isn't big enough to hold the entire header value */
344 while (c
= getc(fp
), c
!= EOF
&& c
!= '\n')
351 static FILE* prepare_ftp_session(FILE **dfpp
, struct host_info
*target
, len_and_sockaddr
*lsa
)
358 target
->user
= xstrdup("anonymous:busybox@");
360 sfp
= open_socket(lsa
);
361 if (ftpcmd(NULL
, NULL
, sfp
) != 220)
362 bb_error_msg_and_die("%s", sanitize_string(G
.wget_buf
+ 4));
365 * Splitting username:password pair,
368 str
= strchr(target
->user
, ':');
371 switch (ftpcmd("USER ", target
->user
, sfp
)) {
375 if (ftpcmd("PASS ", str
, sfp
) == 230)
377 /* fall through (failed login) */
379 bb_error_msg_and_die("ftp login: %s", sanitize_string(G
.wget_buf
+ 4));
382 ftpcmd("TYPE I", NULL
, sfp
);
387 if (ftpcmd("SIZE ", target
->path
, sfp
) == 213) {
388 G
.content_len
= BB_STRTOOFF(G
.wget_buf
+ 4, NULL
, 10);
389 if (G
.content_len
< 0 || errno
) {
390 bb_error_msg_and_die("SIZE value is garbage");
396 * Entering passive mode
398 if (ftpcmd("PASV", NULL
, sfp
) != 227) {
400 bb_error_msg_and_die("bad response to %s: %s", "PASV", sanitize_string(G
.wget_buf
));
402 // Response is "227 garbageN1,N2,N3,N4,P1,P2[)garbage]
403 // Server's IP is N1.N2.N3.N4 (we ignore it)
404 // Server's port for data connection is P1*256+P2
405 str
= strrchr(G
.wget_buf
, ')');
406 if (str
) str
[0] = '\0';
407 str
= strrchr(G
.wget_buf
, ',');
408 if (!str
) goto pasv_error
;
409 port
= xatou_range(str
+1, 0, 255);
411 str
= strrchr(G
.wget_buf
, ',');
412 if (!str
) goto pasv_error
;
413 port
+= xatou_range(str
+1, 0, 255) * 256;
414 set_nport(&lsa
->u
.sa
, htons(port
));
416 *dfpp
= open_socket(lsa
);
419 sprintf(G
.wget_buf
, "REST %"OFF_FMT
"u", G
.beg_range
);
420 if (ftpcmd(G
.wget_buf
, NULL
, sfp
) == 350)
421 G
.content_len
-= G
.beg_range
;
424 if (ftpcmd("RETR ", target
->path
, sfp
) > 150)
425 bb_error_msg_and_die("bad response to %s: %s", "RETR", sanitize_string(G
.wget_buf
));
430 static void NOINLINE
retrieve_file_data(FILE *dfp
)
432 #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT
433 # if ENABLE_FEATURE_WGET_TIMEOUT
436 struct pollfd polldata
;
438 polldata
.fd
= fileno(dfp
);
439 polldata
.events
= POLLIN
| POLLPRI
;
441 progress_meter(PROGRESS_START
);
446 /* Loops only if chunked */
449 #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT
450 /* Must use nonblocking I/O, otherwise fread will loop
451 * and *block* until it reads full buffer,
452 * which messes up progress bar and/or timeout logic.
453 * Because of nonblocking I/O, we need to dance
454 * very carefully around EAGAIN. See explanation at
457 ndelay_on(polldata
.fd
);
463 rdsz
= sizeof(G
.wget_buf
);
465 if (G
.content_len
< (off_t
)sizeof(G
.wget_buf
)) {
466 if ((int)G
.content_len
<= 0)
468 rdsz
= (unsigned)G
.content_len
;
472 #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT
473 # if ENABLE_FEATURE_WGET_TIMEOUT
474 second_cnt
= G
.timeout_seconds
;
477 if (safe_poll(&polldata
, 1, 1000) != 0)
478 break; /* error, EOF, or data is available */
479 # if ENABLE_FEATURE_WGET_TIMEOUT
480 if (second_cnt
!= 0 && --second_cnt
== 0) {
481 progress_meter(PROGRESS_END
);
482 bb_error_msg_and_die("download timed out");
485 /* Needed for "stalled" indicator */
486 progress_meter(PROGRESS_BUMP
);
489 /* fread internally uses read loop, which in our case
490 * is usually exited when we get EAGAIN.
491 * In this case, libc sets error marker on the stream.
492 * Need to clear it before next fread to avoid possible
493 * rare false positive ferror below. Rare because usually
494 * fread gets more than zero bytes, and we don't fall
495 * into if (n <= 0) ...
500 n
= fread(G
.wget_buf
, 1, rdsz
, dfp
);
502 * If error occurs, or EOF is reached, the return value
503 * is a short item count (or zero).
504 * fread does not distinguish between EOF and error.
507 #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT
508 if (errno
== EAGAIN
) /* poll lied, there is no data? */
512 bb_perror_msg_and_die(bb_msg_read_error
);
513 break; /* EOF, not error */
516 xwrite(G
.output_fd
, G
.wget_buf
, n
);
518 #if ENABLE_FEATURE_WGET_STATUSBAR
520 progress_meter(PROGRESS_BUMP
);
524 if (G
.content_len
== 0)
528 #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT
530 ndelay_off(polldata
.fd
); /* else fgets can get very unhappy */
535 fgets_and_trim(dfp
); /* Eat empty line */
538 G
.content_len
= STRTOOFF(G
.wget_buf
, NULL
, 16);
539 /* FIXME: error check? */
540 if (G
.content_len
== 0)
541 break; /* all done! */
545 /* Draw full bar and free its resources */
546 G
.chunked
= 0; /* makes it show 100% even for chunked download */
547 G
.got_clen
= 1; /* makes it show 100% even for download of (formerly) unknown size */
548 progress_meter(PROGRESS_END
);
551 static void download_one_url(const char *url
)
553 bool use_proxy
; /* Use proxies if env vars are set */
555 len_and_sockaddr
*lsa
;
556 FILE *sfp
; /* socket to web/ftp server */
557 FILE *dfp
; /* socket to ftp server (data) */
559 char *fname_out_alloc
;
560 char *redirected_path
= NULL
;
561 struct host_info server
;
562 struct host_info target
;
564 server
.allocated
= NULL
;
565 target
.allocated
= NULL
;
569 parse_url(url
, &target
);
571 /* Use the proxy if necessary */
572 use_proxy
= (strcmp(G
.proxy_flag
, "off") != 0);
574 proxy
= getenv(target
.is_ftp
? "ftp_proxy" : "http_proxy");
575 use_proxy
= (proxy
&& proxy
[0]);
577 parse_url(proxy
, &server
);
580 server
.port
= target
.port
;
581 if (ENABLE_FEATURE_IPV6
) {
582 //free(server.allocated); - can't be non-NULL
583 server
.host
= server
.allocated
= xstrdup(target
.host
);
585 server
.host
= target
.host
;
589 if (ENABLE_FEATURE_IPV6
)
590 strip_ipv6_scope_id(target
.host
);
592 /* If there was no -O FILE, guess output filename */
593 fname_out_alloc
= NULL
;
594 if (!(option_mask32
& WGET_OPT_OUTNAME
)) {
595 G
.fname_out
= bb_get_last_path_component_nostrip(target
.path
);
596 /* handle "wget http://kernel.org//" */
597 if (G
.fname_out
[0] == '/' || !G
.fname_out
[0])
598 G
.fname_out
= (char*)"index.html";
599 /* -P DIR is considered only if there was no -O FILE */
602 G
.fname_out
= fname_out_alloc
= concat_path_file(G
.dir_prefix
, G
.fname_out
);
604 /* redirects may free target.path later, need to make a copy */
605 G
.fname_out
= fname_out_alloc
= xstrdup(G
.fname_out
);
609 #if ENABLE_FEATURE_WGET_STATUSBAR
610 G
.curfile
= bb_get_last_path_component_nostrip(G
.fname_out
);
613 /* Determine where to start transfer */
615 if (option_mask32
& WGET_OPT_CONTINUE
) {
616 G
.output_fd
= open(G
.fname_out
, O_WRONLY
);
617 if (G
.output_fd
>= 0) {
618 G
.beg_range
= xlseek(G
.output_fd
, 0, SEEK_END
);
620 /* File doesn't exist. We do not create file here yet.
621 * We are not sure it exists on remote side */
626 lsa
= xhost2sockaddr(server
.host
, server
.port
);
627 if (!(option_mask32
& WGET_OPT_QUIET
)) {
628 char *s
= xmalloc_sockaddr2dotted(&lsa
->u
.sa
);
629 fprintf(stderr
, "Connecting to %s (%s)\n", server
.host
, s
);
633 /*G.content_len = 0; - redundant, got_clen = 0 is enough */
636 if (use_proxy
|| !target
.is_ftp
) {
644 /* Open socket to http server */
645 sfp
= open_socket(lsa
);
647 /* Send HTTP request */
649 fprintf(sfp
, "GET %stp://%s/%s HTTP/1.1\r\n",
650 target
.is_ftp
? "f" : "ht", target
.host
,
653 if (option_mask32
& WGET_OPT_POST_DATA
)
654 fprintf(sfp
, "POST /%s HTTP/1.1\r\n", target
.path
);
656 fprintf(sfp
, "GET /%s HTTP/1.1\r\n", target
.path
);
659 fprintf(sfp
, "Host: %s\r\nUser-Agent: %s\r\n",
660 target
.host
, G
.user_agent
);
662 /* Ask server to close the connection as soon as we are done
663 * (IOW: we do not intend to send more requests)
665 fprintf(sfp
, "Connection: close\r\n");
667 #if ENABLE_FEATURE_WGET_AUTHENTICATION
669 fprintf(sfp
, "Proxy-Authorization: Basic %s\r\n"+6,
670 base64enc(target
.user
));
672 if (use_proxy
&& server
.user
) {
673 fprintf(sfp
, "Proxy-Authorization: Basic %s\r\n",
674 base64enc(server
.user
));
679 fprintf(sfp
, "Range: bytes=%"OFF_FMT
"u-\r\n", G
.beg_range
);
681 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
683 fputs(G
.extra_headers
, sfp
);
685 if (option_mask32
& WGET_OPT_POST_DATA
) {
687 "Content-Type: application/x-www-form-urlencoded\r\n"
688 "Content-Length: %u\r\n"
691 (int) strlen(G
.post_data
), G
.post_data
696 fprintf(sfp
, "\r\n");
702 * Retrieve HTTP response line and check for "200" status code.
708 str
= skip_non_whitespace(str
);
709 str
= skip_whitespace(str
);
710 // FIXME: no error check
711 // xatou wouldn't work: "200 OK"
716 while (gethdr(sfp
) != NULL
)
717 /* eat all remaining headers */;
721 Response 204 doesn't say "null file", it says "metadata
722 has changed but data didn't":
724 "10.2.5 204 No Content
725 The server has fulfilled the request but does not need to return
726 an entity-body, and might want to return updated metainformation.
727 The response MAY include new or updated metainformation in the form
728 of entity-headers, which if present SHOULD be associated with
729 the requested variant.
731 If the client is a user agent, it SHOULD NOT change its document
732 view from that which caused the request to be sent. This response
733 is primarily intended to allow input for actions to take place
734 without causing a change to the user agent's active document view,
735 although any new or updated metainformation SHOULD be applied
736 to the document currently in the user agent's active view.
738 The 204 response MUST NOT include a message-body, and thus
739 is always terminated by the first empty line after the header fields."
741 However, in real world it was observed that some web servers
742 (e.g. Boa/0.94.14rc21) simply use code 204 when file size is zero.
746 case 300: /* redirection */
756 bb_error_msg_and_die("server returned error: %s", sanitize_string(G
.wget_buf
));
760 * Retrieve HTTP headers.
762 while ((str
= gethdr(sfp
)) != NULL
) {
763 static const char keywords
[] ALIGN1
=
764 "content-length\0""transfer-encoding\0""location\0";
766 KEY_content_length
= 1, KEY_transfer_encoding
, KEY_location
770 /* gethdr converted "FOO:" string to lowercase */
772 /* strip trailing whitespace */
773 char *s
= strchrnul(str
, '\0') - 1;
774 while (s
>= str
&& (*s
== ' ' || *s
== '\t')) {
778 key
= index_in_strings(keywords
, G
.wget_buf
) + 1;
779 if (key
== KEY_content_length
) {
780 G
.content_len
= BB_STRTOOFF(str
, NULL
, 10);
781 if (G
.content_len
< 0 || errno
) {
782 bb_error_msg_and_die("content-length %s is garbage", sanitize_string(str
));
787 if (key
== KEY_transfer_encoding
) {
788 if (strcmp(str_tolower(str
), "chunked") != 0)
789 bb_error_msg_and_die("transfer encoding '%s' is not supported", sanitize_string(str
));
792 if (key
== KEY_location
&& status
>= 300) {
793 if (--redir_limit
== 0)
794 bb_error_msg_and_die("too many redirections");
797 free(redirected_path
);
798 target
.path
= redirected_path
= xstrdup(str
+1);
799 /* lsa stays the same: it's on the same server */
801 parse_url(str
, &target
);
803 free(server
.allocated
);
804 server
.allocated
= NULL
;
805 server
.host
= target
.host
;
806 /* strip_ipv6_scope_id(target.host); - no! */
807 /* we assume remote never gives us IPv6 addr with scope id */
808 server
.port
= target
.port
;
811 } /* else: lsa stays the same: we use proxy */
813 goto establish_session
;
816 // if (status >= 300)
817 // bb_error_msg_and_die("bad redirection (no Location: header from server)");
819 /* For HTTP, data is pumped over the same connection */
826 sfp
= prepare_ftp_session(&dfp
, &target
, lsa
);
831 if (!(option_mask32
& WGET_OPT_SPIDER
)) {
833 G
.output_fd
= xopen(G
.fname_out
, G
.o_flags
);
834 retrieve_file_data(dfp
);
835 if (!(option_mask32
& WGET_OPT_OUTNAME
)) {
842 /* It's ftp. Close data connection properly */
844 if (ftpcmd(NULL
, NULL
, sfp
) != 226)
845 bb_error_msg_and_die("ftp error: %s", sanitize_string(G
.wget_buf
+ 4));
846 /* ftpcmd("QUIT", NULL, sfp); - why bother? */
850 free(server
.allocated
);
851 free(target
.allocated
);
852 free(fname_out_alloc
);
853 free(redirected_path
);
856 int wget_main(int argc
, char **argv
) MAIN_EXTERNALLY_VISIBLE
;
857 int wget_main(int argc UNUSED_PARAM
, char **argv
)
859 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
860 static const char wget_longopts
[] ALIGN1
=
861 /* name, has_arg, val */
862 "continue\0" No_argument
"c"
863 //FIXME: -s isn't --spider, it's --save-headers!
864 "spider\0" No_argument
"s"
865 "quiet\0" No_argument
"q"
866 "output-document\0" Required_argument
"O"
867 "directory-prefix\0" Required_argument
"P"
868 "proxy\0" Required_argument
"Y"
869 "user-agent\0" Required_argument
"U"
870 #if ENABLE_FEATURE_WGET_TIMEOUT
871 "timeout\0" Required_argument
"T"
874 // "tries\0" Required_argument "t"
875 /* Ignored (we always use PASV): */
876 "passive-ftp\0" No_argument
"\xff"
877 "header\0" Required_argument
"\xfe"
878 "post-data\0" Required_argument
"\xfd"
879 /* Ignored (we don't do ssl) */
880 "no-check-certificate\0" No_argument
"\xfc"
884 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
885 llist_t
*headers_llist
= NULL
;
890 IF_FEATURE_WGET_TIMEOUT(G
.timeout_seconds
= 900;)
891 G
.proxy_flag
= "on"; /* use proxies if env vars are set */
892 G
.user_agent
= "Wget"; /* "User-Agent" header field */
894 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
895 applet_long_options
= wget_longopts
;
897 opt_complementary
= "-1" IF_FEATURE_WGET_TIMEOUT(":T+") IF_FEATURE_WGET_LONG_OPTIONS(":\xfe::");
898 getopt32(argv
, "csqO:P:Y:U:T:" /*ignored:*/ "t:",
899 &G
.fname_out
, &G
.dir_prefix
,
900 &G
.proxy_flag
, &G
.user_agent
,
901 IF_FEATURE_WGET_TIMEOUT(&G
.timeout_seconds
) IF_NOT_FEATURE_WGET_TIMEOUT(NULL
),
902 NULL
/* -t RETRIES */
903 IF_FEATURE_WGET_LONG_OPTIONS(, &headers_llist
)
904 IF_FEATURE_WGET_LONG_OPTIONS(, &G
.post_data
)
908 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
912 llist_t
*ll
= headers_llist
;
914 size
+= strlen(ll
->data
) + 2;
917 G
.extra_headers
= cp
= xmalloc(size
);
918 while (headers_llist
) {
919 cp
+= sprintf(cp
, "%s\r\n", (char*)llist_pop(&headers_llist
));
925 G
.o_flags
= O_WRONLY
| O_CREAT
| O_TRUNC
| O_EXCL
;
926 if (G
.fname_out
) { /* -O FILE ? */
927 if (LONE_DASH(G
.fname_out
)) { /* -O - ? */
929 option_mask32
&= ~WGET_OPT_CONTINUE
;
931 /* compat with wget: -O FILE can overwrite */
932 G
.o_flags
= O_WRONLY
| O_CREAT
| O_TRUNC
;
936 download_one_url(*argv
++);
938 if (G
.output_fd
>= 0)